This post in 30 seconds.
- The problem: most unnecessary visits were decided in the ten seconds after someone read a one-line ticket. The dispatcher had a sentence, not a view of the fault.
- The seven: no fault found, user error, settings and consumables, wrong part, wrong skill, access problems, and jobs that belong to someone else. Each one looks different on the ticket and each one shows up on a phone camera.
- The fix: a guided capture that asks for the identity, the display, the connections and the surroundings, in that order, so the desk can close, dispatch prepared, or redirect.
- Who this is for: service managers, dispatch leads and technical support managers at equipment makers, dealers, installers, utilities and maintenance providers.
A technician once drove two hours to a till that "wouldn't take cards". The phone cable was out of the wall. He pushed it back in, the till worked, and the whole day went on the invoice as a completed job.
That story sits on a technician forum from years ago, and every service manager reading it can name their own version from last week. The unit that "wasn't working" because the breaker had tripped, the boiler on "no heat" with the thermostat set to holiday mode, the pump that needed a part nobody had on the van.
If you run dispatch for an equipment dealer, an installer network or a utility, you already know which jobs cost the most. They are the ones that should never have gone on the board, and the ones that went on the board unprepared. This post is about catching both before a van moves, without a live video appointment and without sensors on the installed base.
Most service desks decide on dispatch from a sentence the customer typed, and the desk is right to feel uneasy about it. We built Venta Capture, VentaVid's guided capture platform, so the person already standing at the fault can show it to you before you commit a technician. Start for free and send a test link to your own phone before the next planning meeting.
In this post:
Which service visits are actually unnecessary
"Unnecessary" hides two different failures. In one, the visit should never have happened; in the other, it was needed but failed, so a second one followed. Both show up as waste, and they need different fixes.
The size of the first group is better documented than most people expect. Aquant's 2025 Field Service Benchmark, built on more than 600,000 technician records, found 14% of onsite visits were unnecessary, with avoidable dispatch running at 3% for top performers and 24% for the bottom fifth. For the no-fault-found slice specifically, Cranfield University's Dr John Ahmet Erkoyuncu puts the estimate around 20% of B2B service calls, and adds that companies would not share their real figures with his researchers.
What a truck roll costs is covered in our post on reducing truck rolls, and the full bill for a wasted one is in what an unnecessary field service visit costs. This post stays on the operations question: which visits, and how do you catch them.
The useful move is to stop treating "unnecessary" as one bucket and start tagging every wasted visit with the reason it was wasted. Seven reasons cover almost everything a service desk sees.
Three of these (no fault found, user error, settings and consumables) usually end with no visit at all. Three more (wrong part, wrong skill, access) still need a technician, but a technician who arrives with the right part, the right ticket and a key. The last one belongs to somebody else and should leave your board within the hour.
The second group is the one that quietly doubles your cost. Aquant's same benchmark found that a failed first visit adds two more visits on average and stretches resolution by 14 days. A repeat visit is unnecessary in exactly the same sense as a no-fault-found trip: somebody made a decision without seeing enough.
Why "send us a photo" doesn't stop the visit
Most desks have already tried the obvious fix. The agent asks the customer to text a picture, the customer sends one, and the dispatch happens anyway.
The photo fails for three reasons, and none of them is the customer's fault.
- Wrong subject. The customer photographs what they think is broken. You needed the model plate, the display, and the connection at the back, and you got a wide shot of a cabinet.
- No context. A close-up of a fault code tells you nothing about the unit it came from, how it is installed, or what is plugged into it. Every one of the seven categories above depends on context.
- Wrong channel. The picture lands on a technician's personal WhatsApp or in a personal inbox, three days before the visit, and never reaches the ticket. The dispatcher decides without it.
Verbal triage has the same defect in a different form. The customer answers the question they think you asked ("is it plugged in?" "yes") and the call-centre script has no way to check. The pre-visit assessment that would have caught it needs eyes, and the only camera on site belongs to the customer.
There is a real cost to the back-and-forth as well. Smart Service reports that chasing the missing information delays triage by four to six hours on average, and that photos on file at intake cut the number of wrong-part-on-the-truck returns.
A photo request without a shot list is an open upload, and open uploads return whatever the customer guessed you wanted. Our post on guided photo capture goes deep on why uploads fail; the short version is that the expertise has to sit in the request, not in the customer. The difference is explained in the glossary under upload versus capture.
If your desk is tagging visits this week and the no-fault-found column is longer than you'd like, book a demo and bring three real tickets. We'll build the capture flow for them on the call.
What a guided capture asks for so the desk can decide
A guided capture turns "send us a photo" into a sequence. The organisation decides what is captured, in what order, with what instruction on screen, and the customer follows it on their own phone from a secure, personal capture link sent by SMS, WhatsApp, email or QR code. No app, no account, and, this is the part that matters for dispatch, no appointment.
The sequence below is the one that catches the seven categories. Adjust it per ticket type; the order is deliberate.
- Identity first. Scan the serial or the asset QR (with a picture of where the sticker sits, and manual typing as fallback). This alone kills most wrong-part dispatches, because the desk now knows the exact model before anyone orders anything.
- The display or indicator. A photo of the panel, the fault code, the status lights. Blank display and dark lights is a power question, not a repair.
- The whole unit and its surroundings. One wide shot. Ventilation, stock stacked against the housing, water on the floor, the plant-room door. This is where access problems and someone-else's-fault jobs reveal themselves.
- Connections and power. Back of the unit, the plug, the isolator, the breaker panel. The till with the cable out of the wall never becomes a ticket.
- A short video of the behaviour. "Start the machine and describe out loud what you see and hear." Intermittent faults, noises and cycling problems are impossible in a photo and obvious in fifteen seconds of video with sound.
- Yes/no questions that branch. "Is the filter indicator lit?" sends the customer to a consumables step. "Has anyone worked on this unit in the last 30 days?" flags the third-party installation.
- Access, at the end. Who will be on site, where the keys are, whether a lift or a permit is needed. Asked only when the earlier steps have not already closed the case.
This is what Venta Capture's field service flow does. The guided video capture runs in the mobile browser, each step carries its own instruction and example photo, and a step can be required or optional. Standard flows come ready-made and your own team adjusts the steps and questions in the builder; flows with branching between paths, routing rules and integrations are set up around your process with us.
What lands on the desk is a case, not a folder of images. The answers to each question, the spoken description transcribed and searchable, timestamps, the device context, and your own reference number, so the case can sit against the work order in whatever system holds it. If a step is missing or unclear, the reviewer requests a retake in one click, and the retake asks the customer why they are recording again.
Because the capture is asynchronous, the desk's review time is the only staff time spent: a hundred open tickets do not need a hundred scheduled video calls. Teams using Venta Capture for support report 23% of customer problems solved without dispatching a technician. Our post on remote diagnostics without sensors covers the diagnostic side in depth; here the point is narrower. The desk can now decide.
One boundary worth stating plainly. The capture itself diagnoses nothing, and it should not: the person with the expertise gets a proper look, and the remote triage decision stays with them.
The four outcomes of a reviewed case
A reviewed case ends one of four ways. Making the four explicit is what turns a capture tool into an operations change, because each outcome has a status, an owner and a next action.
- Close remotely. The desk sends the instruction (reset, refit the filter, switch the mode) and asks for a confirmation capture. The ticket closes without a visit and the evidence stays on the case.
- Dispatch prepared. The visit still happens, but the part is on the van, the right skill is assigned, the access is confirmed, and the technician has seen the fault before leaving the depot. This is where first-visit resolution moves.
- Redirect. The case shows an upstream fault or a third-party installation. It leaves your board with the evidence attached, which tends to end the argument about whose problem it was.
- Request a retake. The wide shot is missing, the video has no sound, the plate is unreadable. One click, a reason logged, and the customer captures again when they are next at the unit.
Remote-first is not remote-only, and the honest promise is a better-informed visit, not a promise of none. The visits that still roll are the ones where the technician needs to be there, and they arrive with a plan instead of a sentence.
The completion side of the same loop, where the technician captures the finished job before leaving, is covered in proof of work in field service. Same infrastructure, opposite end of the visit.
What to measure, and how to measure it honestly
A visit that did not happen is not in your system. That is the whole measurement problem, and most "we cut visits by X%" claims quietly ignore it.
Four numbers matter. Two are about visits that never happened, two are about visits that happened and worked.
Here is a method that holds up when someone senior asks how you got the number.
- Run a baseline month first. Before the capture link goes out, tag every dispatched ticket with one of the seven categories and record dispatches per 100 tickets by ticket type. Without this you have nothing to compare against.
- Use a 30-day window everywhere. Aquant's analysis of what first-time fix rate can't tell you shows that 7- and 14-day windows produce false-positive fix rates, and that a big gap between the 7-day and 30-day figure means the team is chasing the number rather than the customer.
- Only count a remote close after the window passes. A ticket closed from the desk on Monday that turns into a visit on the 20th was not avoided. It was delayed.
- Charge the review time and the retakes against the saving. Every capture takes minutes at the desk; some need a second capture. Count that time. A saving that ignores its own cost will not survive the finance review.
- Keep the category tags running. A single headline number hides the useful part: seeing that no-fault-found dropped, wrong-part dropped, and access problems did not, which tells you which step of the flow needs work.
For context, Aquant's 2026 benchmark across 161 organisations puts the industry first-time fix rate at 77% (88% for top performers, 60% for the bottom), median time to resolution at 4.5 days, and finds that one in five cases could have been resolved remotely. TSIA reports a median of 87% of on-site incidents resolved in one visit. And a ServiceMax survey cited by Salesforce found 76% of customers dissatisfied when the first visit does not resolve the issue, which is the number to quote when someone asks why repeat visits matter beyond cost.
Track all four, because a remote resolution rate that rises while the repeat-visit rate also rises means the desk is closing tickets it should have dispatched. The Venta Capture support figures (79% fewer field service visits required, and four out of five issues closed the same day) were measured the way this section describes: against what happened before, over a full window. Expect your own number to depend on your ticket mix.
If you are still choosing a tool category, our remote inspection software buyer's guide separates live video, sensor-based diagnostics and asynchronous capture, and says which fits which volume.
Rolling it out without a project team
The mistake is to start with every ticket type. Start with one.
- Pick the ticket type with the longest no-fault-found column from your baseline month. Usually it is the generic "not working" ticket for one product family.
- Build one flow with the seven steps above, trimmed to that product. Use the example photo per step; customers copy what they see.
- Send it from one route (one product line, one region, one dispatch desk) so the comparison stays clean. The request-by-link step can be pasted into the existing ticket confirmation email, so nothing changes for the agent except one link.
- Review daily for a month, tag the outcome of every case with one of the four outcomes, then compare against the baseline with the 30-day rule.
Venta Capture is free to start, with no credit card, and the first standard flow is usually live in about ten minutes. If you get stuck, book a free setup call and we build the first flow together. Start for free and send the first link to yourself.
To put a value on each avoided trip, use the cost of an unnecessary field service visit. The intake that makes the decision possible is in customer video intake.
Frequently asked questions
What counts as an unnecessary service visit?
Any dispatch where the technician's presence did not change the outcome: no fault found, a user or settings issue that an instruction would have fixed, or a job that belonged to another party. Most teams also count the failed first visit, since the second trip only exists because the first one was unprepared.
How many service visits are unnecessary?
Aquant's 2025 benchmark found 14% of onsite visits were unnecessary, with avoidable dispatch ranging from 3% at top performers to 24% at the bottom fifth. Estimates for no-fault-found alone sit around 20% of B2B service calls, though companies rarely publish their own figure.
Can we reduce unnecessary visits without IoT sensors?
Yes. Sensors only cover connected equipment, which is a minority of most installed bases. A guided capture uses the phone of the person already on site, so it works on a twenty-year-old boiler as well as on a connected machine.
Why not just ask the customer to send a photo?
Because an open photo request returns whatever the customer guessed you wanted, usually without the model plate, the connections or the surroundings. A guided capture specifies each required view in order, asks branching questions, and lets the desk request a retake when something is missing.
Do customers need an app to do a guided capture?
No. The customer opens a secure, personal link in the phone's browser, follows the steps, and submits. There is no app to install and no account to create, which is the main reason completion rates hold up.
How do we measure visits avoided if the visit never happened?
Run a baseline month with every dispatch tagged by reason, then compare dispatches per 100 tickets for the same ticket type after the capture link is in use. Only count a remote close once 30 days have passed without a visit, and subtract the desk's review time from the saving.
What is a good repeat-visit rate?
Lower is better, and the honest benchmark depends on your window. Aquant's 2026 data puts the industry first-time fix rate at 77% with top performers at 88%, both measured over 30 days, so a repeat-visit rate above 20% on a 30-day window means one job in five is coming back.
Talk to us
If a third of your board this week was a visit nobody needed, the fastest check is to send one capture link to one customer and watch what comes back. Most service managers know within a day whether the ticket text was telling them the truth.
What it takes to start.
- A free plan, no credit card, so you can build a flow and run a real ticket through it before anyone signs anything.
- Live in about ten minutes. Standard flows come ready-made; adjust the steps for your product and send.
- Stuck? Book a free setup call and we build your first flow together.
A real person from the VentaVid team runs these demos, on your own ticket types.
Start for free or book a demo and bring three real tickets.